You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@lucene.apache.org by GitBox <gi...@apache.org> on 2022/09/04 12:18:18 UTC

[GitHub] [lucene] msokolov commented on pull request #11743: LUCENE-10592 Better estimate memory for HNSW graph

msokolov commented on PR #11743:
URL: https://github.com/apache/lucene/pull/11743#issuecomment-1236325605

   I ran a test using KnnGraphTester, setting IndexWriterConfig max buffer size to 250MB. I indexed 1M 256-dim docs (1K each, so 1GB in total) Results with 9.3:
   
   ```
   4 segments
   recall  latency nDoc    fanout  maxConn beamWidth       visited index ms
   0.978    2.36   1000000 50      32      64      3977    476926
   ```
   
   With 9.x, 7 segments got created
   
   ```
   recall  latency nDoc    fanout  maxConn beamWidth       visited index ms
   0.983    4.05   1000000 50      32      64      60      522040  1.00    post-filter
   ```
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscribe@lucene.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@lucene.apache.org
For additional commands, e-mail: issues-help@lucene.apache.org